Attempts in Brussels to shed more light on a practice blamed for exacerbating the pre-crisis credit boom are falling short. Europe’s finance watchdog must act to help protect investors from murky “ratings shopping”.
Since 2013 the EU parliament has required credit rating agencies to disclose publicly each initial approach for a rating from an issuer of structured bonds; it also requires every issue of structured bonds to receive a final rating from two agencies.
The objective is to restrict ratings shopping, where issuers contact several agencies to find the one likely to give their bonds the highest rating. Our study shows, however, that the disclosure system is failing: a normal investor has little chance of finding out, in a timely manner, how many credit rating agencies receive an initial approach.
The European Securities and Markets Authority should redesign its system to make it useful to investors in three areas.
First, bond investors are understandably confused by the labels used in agency disclosures. For initial approaches, the big three use request for an “initial review” (Standard & Poor’s), “preliminary review” (Moody’s) and “rating action” (Fitch). Esma should mandate a uniform terminology for disclosures.
Second, there is no easy way for investors to ascertain whether an issuer has made initial approaches to multiple agencies. Instead, they must search for disclosures by the same issuer with S&P, Moody’s and Fitch, as well as smaller agencies. Such checking can take many hours because disclosures are often incomplete and each website has its own format. Esma’s repository of agency disclosures is limited to final ratings. To help investors, the authority should include all initial approaches by an EU issuer for each structured bond. It could easily consolidate disclosures on all initial approaches since they are already submitted by the agencies to Esma.
Third, public disclosures on initial approaches should be accessible to investors as they decide whether to buy bonds. This decision usually occurs when the issuer disseminates a preliminary offering document and makes investor presentations based on a provisional rating. The final rating is usually published at the close of the bond offering.
Yet disclosures on initial approaches are often posted on agency websites only after the publication of the final ratings. As a result, bond investors cannot generally see disclosures in a timely manner. Instead some entity — Esma, the agencies or the issuer — should be required to post disclosures at the same time as investors receive a preliminary offering document.
These severe deficiencies in disclosures about initial approaches meant we could only reach limited conclusions about the frequency of ratings shopping by structured finance issuers in the EU.
For 2016, we did our best to gather all disclosures of initial approaches by EU issuers to the big three agencies. We then searched the databases of Esma and the agencies to determine if these approaches led to final credit ratings by the same agencies for these same deals.
We found that 94 per cent of initial approaches to Fitch led to final ratings, with Moody’s scoring 96 per cent. By contrast, only 85 per cent of initial approaches to S&P led to final ratings. These statistics suggest that S&P might have been less ready to indicate relatively high ratings on initial approaches than its two rivals.
In short, the current system for disclosing initial approaches by issuers of structured bonds is poorly designed to constrain ratings shopping.
To provide an effective deterrent, Esma should require the use of the same terminology by all agencies and timely disclosures of initial approaches to investors.
Most importantly, Esma should establish a central and accessible system of public disclosures on both initial approaches and final ratings by each EU issuer of a structured bond.
Robert Pozen is a senior lecturer at the MIT Sloan School of Management and a senior fellow at the Brookings Institution. Sichen Chao, an MIT undergraduate, collected and analysed the data for this article